IN PROGRESS2025-26

Tesseract

3D Semantic Knowledge Visualization

A 3D navigable knowledge space where ideas exist as objects in semantic geometry. Every concept, project, and conversation stored in Anamnesis gets embedded and projected into three-dimensional space using UMAP, so proximity means relatedness and navigation means thinking. Built as the visual exploration layer on top of Anamnesis, Tesseract turns a knowledge graph into a place you can fly through. Clusters emerge naturally from the data: adjacent ideas pull together, distant ones drift apart. The goal is an interface that feels less like search and more like spatial memory.

Connected Systems

Anamnesisknowledge and memory layer
LIVE DEMOopen in new tab

Demo requires login. Open directly

Live instance coming soon. Built on top of the Anamnesis knowledge graph.

Target Specifications

>UMAP projection of all Anamnesis embeddings into 3D space
>Real-time node fly-through: click any concept to open it in Anamnesis
>Cluster auto-labeling: emergent topic regions from semantic density
>Timeline scrubbing: watch the knowledge graph grow over time
>Nucleus integration: spatial map of project context and decisions

Tech Stack

Three.jsReact Three FiberUMAP-jsNext.js 15FastAPINeo4jWeaviatePython 3.11

Tags

CodeUX/UI

Active Sub-Projects

>Anamnesis semantic layer: embedding pipeline and vector store
-Anamnesis semantic layer: embedding pipeline and vector store
-3D force-directed graph renderer (Three.js / React Three Fiber)
-UMAP projection server: Python FastAPI endpoint
-Cluster detection and auto-labeling layer